A Centralised PMO Works Only When Enterprise Coordination Beats Silo Control

Centralisation is not a maturity badge; it is an operating-model choice that should earn its cost by solving coordination problems the organisation cannot solve locally.

A functional project manager often understands the work better than a central office. They know the people, systems, constraints and technical detail. So why centralise anything?

Because enterprise initiatives create problems that functional excellence cannot solve alone.

Projects compete for the same specialists. One business unit’s optimisation creates another unit’s constraint. Investment priorities conflict. Dependencies cross organisational boundaries. Executives receive inconsistent reporting. Subject-matter experts are asked to manage projects while still carrying full operational responsibilities.

A centralised PMO can address these problems. It can also create bureaucracy, distance and a false sense of control.

The decision should therefore be based on coordination economics, not fashion.

The Strategic Context

Accenture’s 2017 practitioner paper on a centralised PMO argues that growing cross-functional complexity can expose limitations in relying on standalone functional project managers. It highlights resource allocation, standardised processes, project oversight and enterprise-wide coordination as potential PMO benefits.

The paper also acknowledges the advantages of functional ownership: subject-matter expertise sits close to the work. The tension is therefore genuine.

Crawford’s 2010 strategic-PMO paper adds a broader enterprise perspective: a strategic PMO can provide an organisational home for linking strategy to programs and projects, portfolio prioritisation, performance information and scarce-resource optimisation.

Neither source proves that every organisation should centralise. Together they identify the conditions under which centralisation can become economically rational.

What Leaders Commonly Misread

The first mistake is to centralise because project performance is poor without diagnosing the cause. If the real issue is weak sponsorship, unrealistic strategy or overloaded capacity, moving project managers into a PMO may not solve it.

The second is to assume standardisation automatically creates efficiency. Standard processes reduce reinvention only when they replace genuinely repeatable work. Poorly designed standards add overhead.

The third is to equate independence with neutrality. A central PMO may be less tied to one function, but it still needs legitimate executive sponsorship and transparent prioritisation.

The fourth is to strip functional teams of ownership. The PMO should not become the place where the business “hands over” change. Benefits still have to be realised in operations.

Related article: When a PMO's Mandate Outgrows Its Place on the Organisation Chart

Reframing the Issue

Ask a more precise question:

Where is the cost of decentralised coordination greater than the cost of centralisation?

Centralisation becomes attractive when several conditions are present:

initiatives cross multiple functions;
scarce specialists are shared;
enterprise priorities compete;
dependencies create systemic risk;
executives need comparable information;
local managers have conflicts between operational and project responsibilities;
repeated project infrastructure is being rebuilt in each business unit;
portfolio decisions require a whole-of-enterprise view.

Conversely, decentralisation may be stronger where work is specialised, local, independent and fast-moving, and where enterprise coordination adds little value.

The optimum model may be federated rather than fully centralised.

The PMO Should Centralise Coordination, Not Expertise

Accenture’s paper is strongest when it argues for combining project-management capability with internal subject-matter expertise.

That suggests a design principle: centralise the mechanism, not all the knowledge.

The PMO can own portfolio visibility, resource conflict resolution, common governance, project-management capability and escalation while functional experts remain embedded in delivery.

This avoids two extremes.

In the first, functional experts manage projects in addition to day jobs, creating overload and local bias.

In the second, a central PMO becomes detached from operational reality and substitutes methodology for subject knowledge.

A federated model often works because it keeps enterprise coordination central while preserving expertise close to the work.

Resource Allocation Is the Strategic Test

Accenture identifies resource allocation as a core PMO function. That deserves elevation.

If the PMO cannot influence which initiatives receive scarce people, it may have visibility without leverage. It can report that five programs all require the same architect, engineer or change leader, but the conflict remains unresolved.

Resource allocation also exposes whether the PMO is connected to portfolio governance. Selecting people based only on availability can perpetuate low-value work. Selecting on strategic priority requires a decision system above individual projects.

Related article: Capacity Is a Strategic Constraint: Match Ambition to What the Organisation Can Absorb

Decision Framework

Use a Centralisation Value Test across six dimensions.

Dimension	Favour centralisation when…	Favour local control when…
Dependencies	Cross-functional effects are high	Work is largely independent
Resources	Scarce capabilities are shared	Teams are dedicated locally
Priorities	Enterprise trade-offs are frequent	Local priorities dominate legitimately
Governance	Comparable decisions and escalation are needed	Context requires highly specialised governance
Capability	Project skills are uneven or duplicated	Strong capability already exists locally
Speed	Coordination delays dominate	Central approval would become the delay

Do not choose the model based on organisational size alone. Complexity, interdependence and strategic coordination matter more.

From Strategy to Execution

Immediate action: map the current cost of decentralisation. Identify duplicate methods, unresolved resource conflicts, contradictory priorities, inconsistent data and cross-functional issues that repeatedly reach executives late.

Medium-term capability: centralise only the services that solve those problems. This may include portfolio reporting, assurance, project-manager capability, resource coordination, common governance and standards. Keep domain expertise and benefit ownership in the business.

Long-term positioning: evolve toward a networked execution model where central visibility and enterprise prioritisation coexist with empowered delivery. The PMO should make the organisation easier to coordinate, not harder to navigate.

Centralisation Should Have an Economic Case

A PMO consumes senior talent, systems, process time and organisational attention. Its business case should therefore include more than improved compliance.

Potential value mechanisms include reduced duplicated effort, better allocation of scarce resources, faster escalation, improved decision quality, fewer conflicting initiatives, stronger project capability and more reliable executive visibility.

Not all benefits will be easily monetised. But they should be observable.

If the PMO cannot explain which coordination costs it is reducing, leaders should question whether centralisation has become self-justifying.

Related article: Portfolio Prioritisation Is Not Ranking: Decide What to Accelerate, Defer and Stop

Signals to Monitor

Watch for functions recreating their own shadow PMOs, project teams bypassing central processes, central reports that executives do not use, and persistent resource conflicts the PMO can see but cannot resolve.

Also monitor whether centralisation reduces or increases decision latency. A PMO that creates one source of truth but adds multiple approval layers may improve data while damaging execution speed.

Positive signals include fewer duplicated governance processes, earlier identification of cross-project conflicts, stronger use of specialist project capability and more explicit enterprise trade-offs.

Questions for the Leadership Team

Which coordination problems are we trying to solve by centralising the PMO?
What expertise must remain inside functions for delivery to stay grounded?
Can the PMO influence scarce-resource allocation, or only report conflicts?
Where would a federated model outperform full centralisation?
Which PMO services create measurable enterprise value and which mainly create internal activity?
Has centralisation improved decision speed as well as visibility?

Closing Perspective

A centralised PMO is valuable when the organisation has an enterprise coordination problem.

It is not valuable merely because centralisation looks mature.

The correct operating model is the one that gives leaders a whole-of-enterprise view while keeping expertise, accountability and decision speed where they create the most value.